smaller ice sheet during the Swedish Middle Weichselian (MIS 3)?
HELENA ALEXANDERSON, TIMOTHY JOHNSEN AND ANDREW S. MURRAY
BOREAS
Alexanderson, H., Johnsen, T. & Murray, A. S. 2010 (April): Re-dating the Pilgrimstad Interstadial with OSL: a warmer climate and a smaller ice sheet during the Swedish Middle Weichselian (MIS 3)? Boreas, Vol. 39, pp. 367–376. 10.1111/j.1502-3885.2009.00130.x. ISSN 0300-9483.Pilgrimstad in central Sweden is an important locality for reconstructing environmental changes during the last glacial period (the Weichselian). Its central location has implications for the Scandinavian Ice Sheet as a whole. The site has been assigned an Early Weichselian age (marine isotope stage (MIS) 5 a/c;474 ka), based on pollen stratigraphic correlations with type sections in continental Europe, but the few absolute dating attempts so far have given uncertain results. We re-excavated the site and collected 10 samples for optically stimulated lumines-cence (OSL) dating from mineral- and organic-rich sediments within the new Pilgrimstad section. Single aliquots of quartz were analysed using a post-IR blue single aliquot regenerative-dose (SAR) protocol. Dose recovery tests were satisfactory and OSL ages are internally consistent. All, except one from an underlying unit that is older, lie in the range 52–36 ka, which places the interstadial sediments in the Middle Weichselian (MIS 3); this is compatible with existing radiocarbon ages, including two measured with accelerator mass spectrometry (AMS). The mean of the OSL ages is 446 ka (n = 9). The OSL ages cannot be assigned to the Early Weichselian for all reasonable adjustments to water content estimates and other parameters. The new ages suggest that climate was relatively mild and that the Scandinavian Ice Sheet was absent or restricted to the mountains for at least parts of MIS 3. These results are supported by other recent studies completed in Fennoscandia.
Helena Alexanderson (e-mail: helena.alexanderson@umb.no), Department of Plant and Environmental Sciences, Norwegian University of Life Sciences, P.O. Box 5003, NO-1432 A˚s, Norway, and Department of Physical Geo-graphy and Quaternary Geology, Stockholm University, SE-106 91 Stockholm, Sweden; Timothy Johnsen (e-mail: timothy.johnsen@geo.su.se), Department of Physical Geography and Quaternary Geology, Stockholm University, SE-106 91 Stockholm, Sweden; Andrew S. Murray (e-mail: anmu@risoe.dtu.dk), Nordic Laboratory for Lumine-scence Dating, Department of Earth Sciences, Aarhus University, Risø DTU, DK-4000 Roskilde, Denmark; received 10th February 2009, accepted 6th October 2009.
Pilgrimstad in central Sweden (Fig. 1) is an important
site for reconstructing environmental changes during
the Weichselian in Scandinavia, since it is situated close
to the former ice divide of the Scandinavian Ice Sheet
and contains subtill organic and minerogenic sediments
(Fig. 2). The site has been investigated and described by
a number of authors over the past
70 years (mainly
Kulling 1945; Fr ¨odin 1954; Lundqvist 1967; Robertsson
1988a, b; Garcı´a Ambrosiani 1990), but the absolute
chronology of the site is still poorly known. In this
study, we present and evaluate results of optically
sti-mulated luminescence (OSL) dating that place the
Pil-grimstad Interstadial in marine isotope stage (MIS) 3.
OSL has been used successfully to date Weichselian
deposits in, for example, Russia (Svendsen et al. 2004;
Thomas et al. 2006), Greenland (Hansen et al. 1999;
Adrielsson & Alexanderson 2005), the Himalayas
(Spen-cer & Owen 2004) and New Zealand (Preusser et al.
2005), but in Scandinavia results have been of varying
quality (Kjær et al. 2006; Alexanderson & Murray 2007;
Lagerb ¨ack 2007; Houmark-Nielsen 2008). Because of
this, and because our results are controversial with respect
to previous age determinations of the site, in this article
we focus on the methodology and reliability of the OSL
ages, while the palaeoglaciological and palaeoecological
implications of the dates will be discussed elsewhere.
The Pilgrimstad site
Kulling (1945) recognized three separate series of sand
and gravel within the subtill sediments at Pilgrimstad.
The lowermost series was interpreted as deposited in a
proglacial sub-aquatic environment, while the upper
two series represent a transition from glacifluvial to
flu-vial to lacustrine deposition and contain fine-grained
minerogenic and organic material (Lundqvist 1967). A
well-sorted sandy bed within the lacustrine sediments
has been interpreted as an aeolian deposit (Robertsson
1988a, b). The sediments were exposed at the surface for
some time before the most recent ice advance. Detailed
descriptions of the stratigraphic units and a review of
interpretations are available in Lundqvist (1967).
The palaeoecological interpretation as a cool,
sub-arctic–arctic environment is based on several proxies,
mainly from the organic beds. According to pollen and
coleoptera, open herb–shrub vegetation was followed
by a forest border setting during a climatic optimum
(Robertsson
1988a, b).
This
warm
interval
was
succeeded by a colder climatic phase with periglacial
conditions and a subsequent phase of herb–shrub
vegetation (Robertsson 1988b). Both diatoms and
insects record deposition in a nutrient-rich lake
(Lund-qvist 1967; Robertsson 1986). The insect fauna also
indicates a cool, arctic–subarctic climate, consistent
with the finds of mammoth (Mammuthus primigenius),
reindeer (Rangifer tarandus) and elk (Alces alces)
(Lundqvist 1967).
The site has been assigned an Early Weichselian
(MIS 5 a/c) age based on a correlation of the
palaeo-environmental reconstruction with type sections in
continental Europe and represents the type section for
the local Pilgrimstad Interstadial (Kulling 1967; cf. also
J ¨amtland Interstadial; Lundqvist 1967; Lundqvist &
Miller 1992). Robertsson (1988a, b), for example,
cor-related the organic-rich beds with one or possibly two
Early Weichselian interstadials (Brørup and/or
Odder-ade) depending on how the climatic cooling in the
mid-dle of the section is interpreted – as a phase within one
interstadial or as separating two different interstadials.
So far, there have been few absolute dating attempts
and these have given uncertain results. According to a
recent evaluation of all radiocarbon dates from
Pil-grimstad, 10 samples that are acceptable from a quality
perspective range between 59 and 46 cal. ka BP in age
(Wohlfarth 2009) (Fig. 3). Moreover, a single
thermo-luminescence measurement resulted in 60 ka for the
sandy unit within the organic beds (Garcı´a Ambrosiani
1990) and one U/Th measurement provided an age of
384 ka (Heijnis in Robertsson & Garcı´a Ambrosiani
1992). Recently, Ukkonen et al. (2007) also re-dated a
0 100 200 400Kilometers 40°E 30°E 20°E 20°E 10°E 10°E 0° 70°N 65°N 55°N 65°N 60°N 55°N SWEDEN FINLAND NOR WAY DENMARK EST ONIA LITHU ANIA LATVIA RUSSIA Pilgrimstad Sokli Ruunaa Hitura
0-600 m a.sl. >600 m a.s.l. mammoth 38-32 ka LGM margin
Skåne
Fig. 1. Location map of Pilgrimstad. Mam-moth locations from Ukkonen et al. (2007), Last Glacial Maximum (LGM) margin from Svend-sen et al. (2004). Other sites that also indicate ice-free conditions during MIS 3 are shown: Sokli (50 ka; Helmens et al. 2007a, b), Ruunna (50–25 ka; Lunkka et al. 2008), Hitura (deglacia-tion 62–55 ka; Salonen et al. 2008) and several sites in Sk ˚ane (39–24 ka; Kjær et al. 2006).
Fig. 2. Photograph of the upper part of the new section at Pilgrim-stad (see Fig. 3 for comparisons). D–H represent the lithological units.
mammoth molar from Pilgrimstad to 34–29 cal. ka BP.
The exact stratigraphic position of the mammoth
re-mains is not clear, but based on Kulling’s (1945) and
Robertsson’s (1988a) stratigraphic descriptions, it
seems to derive from the lower part of the organic beds
analysed by Robertsson (1988a).
These ages are all younger than the age estimate based
on pollen stratigraphy; a possible correlation of
Pilgrim-stad with the Moershoofd InterPilgrim-stadial (46–44 ka BP)
(Behre & van der Plicht 1992) has therefore been
pro-posed but considered less likely (Robertsson 1988a).
Setting
The study site is located near Pilgrimstad in J ¨amtland,
central Sweden (Fig. 1) and is situated in an abandoned
gravel pit located at the edge of a valley floor. Most of the
previously described sections have been mined or
cov-ered by colluvium or fill. A small hill in the southern part
of the pit was the focus of our excavation (Fig. 2). The
top (including the till cover) had been removed
pre-viously, and for a short time the section had been covered
by blast stone (J. Lundqvist, pers. comm. 2007). Thus, we
adjusted the sample depths by adding 1 m, corresponding
to the general till thickness in the area. The position of
the investigated section is 62
157.5
0N, 15
101.1
0E and its
elevation 300 m a.s.l., as measured by a Garmin GPS
Vista C and checked against topographic maps.
The beds in the excavated section were correlated to
the stratigraphies presented by Kulling (1945) and
Ro-bertsson (1988b). RoRo-bertsson’s original section was
si-tuated adjacent to and at right angles to our new
section; her stratigraphy is therefore best comparable to
the new stratigraphy.
Methods
Sampling, preparation and measurement
Fieldwork was conducted in September 2006 and in July
2007. Five OSL samples were collected during each
campaign at the levels marked in Fig. 3 (see also Table 1).
The samples were taken in opaque plastic tubes and
stored in black bags until opened under darkroom
Unit A. Sandy gravel with lenses of sand and silt. Unit H. Yellow sand, well-sorted, homogeneous. Wedge-like structure. Contains rip-ups of gyttja.
Unit C. Organic-rich sand. Compact, brecciated.
Unit B. Varved clay and silt. Unit E. Sandy gyttja, smaller brecciated pieces than in D. Unit G. Laminated and massive sand, deformed in lower part. Yellowish to rusty colour. Unit F. Brown sandy gyttja with occasional pebbles <15 cm. B Sediment description 39.2±2 43±5 45±4 48±8 39±3 38±3 >40 OSL (ka) 52±4 46±3 74±5 49±4 36±3 C (ka BP)
All radiocarbon ages in right column from Wohlfarth (2009) except for 34-29 cal. ka (Ukkonen et al. 2007), TL age from García Ambrosiani (1990). These ages are from other sections at the Pilgrimstad site.
Uncertain stratigraphic position.
60 ka TL A Pilgrimstad section x x x x x x x x x x G F H D E B A H C 344 345 346 347 348 318 319 320 321 322 m 1 SGy SGy SiGyb CSil CoGmm GySb Sm Sm/Sl SiSl Sm SiSm SiSm Sm(ng) NW SE NE SSil ~46 ~48 ~47 ~50 ~52 ~59 ~52 ~55 ~52 ~47
Unit D. Grey sandy silty gyttja. Compact, brecciated.
C Chronology
34-29 Correlated ages
(cal. ka BP)
Fig. 3. A. Sketch of the new section at Pilgrimstad, lithology and position of OSL and14C samples. B. Brief sediment description. C. Results of
conditions at Stockholm University (2006) and the
Nor-wegian University of Life Sciences (2007), where the
in-itial preparation was completed. Final preparation,
including heavy liquid separation (2.62 g/cm
3) to remove
feldspars (081318–22 only), treatment with 10% HCl for
5–30 min, 10% H
2O
2for 15–30 min, 38% HF for
60–120 min and 10% HCl again for 40 min, was done at
the Nordic Laboratory for Luminescence Dating, where
the OSL measurements were also undertaken.
The samples were analysed using large aliquots of
quartz (180–250
mm) on Risø TL/OSL readers
equip-ped with calibrated
90Sr/
90Y beta radiation sources
(dose rate 0.14–0.35 Gy/s), blue (47030 nm; 50 mW/
cm
2) and infrared (880 nm,
100 mW/cm
2) light
sour-ces, and detection was through 7 mm of U340 glass
filter (Bøtter-Jensen et al. 2000). Analyses employed
post-IR blue SAR protocols (Murray & Wintle 2000,
2003; Banerjee et al. 2001) adapted to suit the samples
based on dose recovery and preheat experiments (first
batch: preheat 260
1C for 10 s, cut-heat 2201C; second
batch 240
1/2001C). A relatively high test-dose (50 Gy)
was necessary to obtain a statistically precise test signal,
and 100 s of illumination at 280
1C between cycles
im-proved recuperation (response to zero dose).
We calculated the dose rates from gamma
spectro-metry data (Murray et al. 1987) (Table 2) and included
the cosmic ray contribution (Prescott & Hutton 1994).
Natural and saturated water content was measured
with pF rings (cylinder volumeters) (Table 1). To
ac-count for water content changes through time due
mainly to compaction, especially for the organic-rich
sediments, we applied a simple three-stage model to all
our samples (Table 3). The mean water content
ð wÞ
since time of deposition was then calculated as:
w ¼ w
1t
1þ w
2t
2þ w
3t
3ð1Þ
Two samples were collected for radiocarbon dating.
Small twigs (unknown species) were picked out and
dated by AMS
14C at the Lund University
Radio-carbon Dating Laboratory.
Data analysis
Simple component analysis of the continuous wave
OSL data from some aliquots was undertaken using
SigmaPlot 10.0 based on the parameters and formulas
of Choi et al. (2006). The results of the component
analysis are discussed further below (Fig. 4), but based
on such information from dose recovery measurements
(Fig. 5) we chose channels 1–2 (first 0.16 s) and channels
4–6 (0.32–0.56 s) as peak and background integration
limits for all aliquots. The equivalent doses were then
calculated in Risø Luminescence Analyst 3.24
(ex-ponential curve fitting) and in Microsoft Excel. To be
accepted, aliquots had to pass the following rejection
criteria: recycling ratio within 20% of unity,
recupera-tion
o5%, equivalent dose error o50% and signal
more than 3
s above the background. Decay and
growth curves also had to be regular. Ages were
calcu-lated using the mean and median of the equivalent dose
population of accepted aliquots for each sample, as well
as using the natural and saturated water contents.
We also did a sensitivity analysis to determine
quantitatively which uncertainties have the largest
effect on age. Ages were recalculated after adjustments
were made to each of the parameters in turn, using
reasonable estimates of uncertainty for each parameter.
The estimated uncertainties (1 SD) used in these
cal-culations were: depth below surface
1 m, elevation
50 m, grain size 10%, water content (gamma
and beta)
10%, dose rate gamma 5%, dose rate
beta
5%, internal dose rate 30%, density 10%,
cosmic ray contribution
5% and beta source
cali-bration
2%.
Results
Sedimentology and stratigraphy
We distinguished eight units in the new section at
Pil-grimstad (shown and briefly described in Figs 2 and 3).
Table 1. Pilgrimstad OSL sample properties and settings. Listed in order of sample number. For lithofacies codes and stratigraphic unit numbers, see Fig. 3.
Sample Sampling depth (m) Stratigraphic unit Lithofacies code Water content (weight %) Organic content (%) w1 w2 w3 w 061344 1.2 G Sm 30 29.5 9.1 28 2.2 061345 1.8 H SiSm 30 29.4 5.2 27 3.9 061346 2.4 F SGy 150 41.9 16.9 72 9.1 061347 2.9 H Sm/Sl 40 35.7 8.5 34 2.0 061348 3.1 H Sm/Sl 35 32.8 8.0 31 1.9 081318 0.5 H SiSm 35 32.8 5.5 31 1.4 081319 0.5 G SiSl 40 38.0 4.9 35 1.4 081320 2.1 H Sm 35 34.6 3.8 32 1.2 081321 4.1 C GySb 35 34.2 9.0 32 1.1 081322 6.3 A Sm(ng) 35 32.7 6.2 31 2.7
The sediments have been tectonized, as indicated by
brecciation and deformation structures. Within the
limited exposure available to us, the sediments
never-theless seem to have a pancake stratigraphy, with the
exception of unit H, which cuts units E, F and G.
Unit A in the new section correlates with the
‘lime-stone-rich pebbly gravel series’ of Kulling (1945), while
units B–H correspond to his ‘silt-stratified sand series’.
We interpret these sediments as showing an
environ-mental succession from glacifluvial (unit A) to
glacila-custrine (unit B) to laglacila-custrine (units C, D, E, F) and
back to fluvial or glacifluvial (unit G), in line with
pre-vious interpretations. The sandy unit H might represent
the aeolian sand of Robertsson (1988a, b). However,
the cross-cutting relationship with the other units
in-dicates formation after the deposition of units F and G,
and suggests a possible glacitectonic rather than aeolian
genesis. We therefore interpret unit H as a clastic dyke
formed subglacially during the Late Weichselian ice
advance (cf. Larsen & Mangerud 1992; Linde´n et al.
2008). The sand is thus likely reworked sand from unit
G, and the gyttja clasts are rip-ups from the
surround-ing beds. This has implications for the interpretation of
the organic beds as belonging to one or two events, and
for the ages of unit H, as discussed below.
OSL characteristics and ages
The major sample properties, settings and results are
lis-ted in Tables 1–3 and are shown in Figs 3 and 7. The
quartz from Pilgrimstad is insensitive, which necessitated
careful selection of peak and background channels to
best isolate the fast component (Fig. 4). For the 21
ac-cepted dose recovery experiments (out of 27), the average
proportions of the net component signal to the total net
signal used for dose calculation were 928% (fast),
8
7% (medium) and 0.20.3% (slow). When these
channels were selected for peak and background
integra-tion, the resulting signal was dominated by the fast
component. Doses calculated with this channel selection
gave good dose recovery (1.05
0.04, n = 21) (Fig. 5)
and demonstrated that the SAR protocols used were able
accurately to recover a known dose administered before
any heating. The signals were not close to saturation
(Fig. 6).
Equivalent doses range between 89 and 145 Gy and
dose rates between 2.5 and 3.1 Gy/ka (Table 4). The
nine upper samples are 52–36 ka, while the lowermost is
older (74 ka) (Table 4 and Fig. 7). OSL ages from unit
H average 447 ka (n = 5), from unit G 413 ka
(n = 2) and single ages from unit F and C are 488 ka
and 46
3 ka, respectively. The sensitivity analysis
Table 2. Summary of radionuclide concentrations measured with high-resolution gamma spectrometry on the Pilgrimstad OSL samples. Beta and gamma dose rates refer to dry material; for water contents and final dose rates, see Tables 1 and 3.
Sample 238U (Bq/kg) 226Ra (Bq/kg) 232Th (Bq/kg) 40K (Bq/kg) Beta (Gy/ka) Gamma (Gy/ka)
061344 396 42.10.7 35.40.7 64212 2.180.04 1.250.04 061345 345 42.80.6 34.70.5 6689 2.140.04 1.150.04 061346 17810 100.81.2 450.9 67513 3.190.07 1.840.09 061347 335 49.10.7 370.5 71510 2.380.04 1.370.04 061348 283 48.50.5 370.4 7167 2.360.03 1.360.04 081318 408 40.20.8 34.30.8 66315 2.220.05 1.240.04 081319 276 38.50.7 33.20.7 67012 2.170.04 1.210.03 081320 295 31.90.5 30.70.5 68012 2.150.04 1.140.03 081321 4110 100.31.3 35.10.9 64915 2.500.07 1.650.09 081322 427 74.40.9 30.70.7 61411 2.260.05 1.390.06
Table 3. Water-content modelling to calculate average water content since the time of deposition, accounting for compaction and environ-mental changes. For values, see Table 1.
Assumptions Three-stage hydrological evolution
The sediments have never been drier than at the time of sampling, and the natural water content is a minimum water content. (The samples were taken within a large gravel pit in which the groundwater table has been lowered.)
The saturated water content is the maximum water content for the sediments in their present state. (It is limited by porosity.) The porosity of minerogenic sediments did not change significantly with compaction. Most of the compaction took place early in the sediments’ history due to continued sedimentation and, later, pressure from the ice cover.
1. Lake/fluvial stage before the ice advance (sediments loose and saturated).
(a) Water content w1is the
saturated value rounded up to the nearest 5 or 0 for minerogenic samples and an assumed value of 150% for the organic sample (061346), derived from young samples with similar organic content.
(b) Duration is30% of the time since deposition (t1= 0.3).
2. Ice cover (sediments compacted and saturated).
(a) Water content w2is the
saturated value.
(b) Duration is60% of the time since deposition (t2= 0.6).
3. After deglaciation (sediments compacted and drier). (a) Water content w3is the
natural value.
(b) Duration is10% of the time since deposition (t3= 0.1).
demonstrates that the calculated age is most sensitive
to changes in the equivalent dose, beta source
calibra-tion and dose rate, followed by water content (Fig. 8).
New radiocarbon ages from twigs are 39.22 and
440
14C ka BP (Table 5).
Discussion
OSL and sediments
From a sedimentological perspective, we can identify
two potential problems for OSL-dating at this site: (1)
the glacifluvial deposit (unit A) may be incompletely
bleached and (2) we cannot properly estimate the
ori-ginal mean water content of the relatively organic-rich
unit F.
Incomplete bleaching results in an apparent age
overestimation, and is fairly common in glacial settings
(e.g. Fuchs & Owen 2008). As we have used only large
aliquots for measurements, it is difficult to infer
any-thing about incomplete bleaching from the dose
dis-tribution of sample 081322, which is from unit A. In
combination with the stratigraphic position of unit A
(lowest), we thus consider the OSL age from unit A as
providing a maximum age for the overlying organic
beds. The good correspondence of ages (n = 9) from all
the other beds (derived from various depositional
set-tings) suggests that there are no problems with
in-complete bleaching for those.
Organic-rich deposits tend to be fairly loose at the
time of deposition and may have very high water
con-tent, i.e. up to several hundred percent. With time they
will become compacted due to sediment loading and the
overriding ice sheet; the water content will change
sig-nificantly, so that what is measured today is not
Growth curve using standard data
0.0 0.5 1.0 1.5 2.0 2.5 0 200 400 600 800 1000 1200 Dose (s) Lx /T x
Growth curve using derived fast component data
0.0 0.5 1.0 1.5 2.0 2.5 0 200 400 600 800 1000 1200 Dose (s) Lx /T x 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% A B C 0 1 2 3 4 5
Stimulation time (sec)
% of bulk signal 0 500 1000 1500 2000 2500 3000 3500 4000 Bulk signal natural signal modeled signal fast medium slow background Sample 061347 given dose given dose Sample 061347 Sample 061347 Recycling = 1.13 Dose recovery = 0.90 Recycling = 1.11 Dose recovery = 0.92
Fig. 4. Data derived from the 27 dose recovery experiments allowed selecting peak and background channels that provided the overall best recycling and dose recovery and a dominance of the fast component after background subtraction. A. Signal component analysis of sample 061347 showing the natural signal which was de-convoluted into fast, medium and slow components (left axis), and the natural and modelled (sum) signal (right axis). Also shown are the peak and background portions of the signal used (0–0.16 s and 0.32–0.56 s). Note that only the first 5 s of the total length of the stimulation (40 s) are shown. B. The upper growth curve uses the standard data (bulk signal) with the selected channels for the same aliquot. C. The lower growth curve is produced by using the derived (de-convoluted) fast component signal data only. The insignificant difference between the growth curves indicates that the channel selection effectively isolated the fast component and that the selection of peak and background channels provided the optimum recycling and dose recovery.
0 2 Frequency 4 0-0.1 0.2-0.3 0.4-0.5 0.6-0.7 0.8-0.9 1-1.1 1.2-1.3 1.4-1.5 1.6-1.7 1.8-1.9
Ratio of measured to given dose Mean = 1.05 Std err = 0.04 n = 21
Fig. 5. Histogram of dose recovery tests, excluding aliquots that did not pass rejection criteria. A mean close to unity indicates that the samples can be used for OSL-dating with our analytical protocols.
representative of the mean water content since time of
deposition (cf. Alexanderson et al. 2008). An
under-estimation of the mean water content results in OSL ages
that appear too young, and vice versa. In our case, the
OSL ages from the surrounding minerogenic beds can
provide some constraints, since these sediments do not
suffer to the same degree from water-content variations.
The sandy unit H is the youngest, since it cross-cuts
units E, F and G. The OSL ages from unit H should
thus be minimum ages for the organic deposits.
How-ever, as mentioned above, re-interpretation of the unit
as a clastic dyke with remobilized material implies that
the sand in unit H could be contemporaneous with unit
G and that the OSL ages do not represent the timing of
the formation of the dyke. It also implies that the
se-paration into units E and F is secondary.
Based on the stratigraphic information and on the
OSL ages, we consider the OSL ages from units B to H
as representing one event, while unit A is older. The
OSL ages from units B–H are internally consistent, i.e.
all lie in the range 52–36 ka, with a mean of 446 ka;
this places the interstadial sediments in the Middle
Weichselian (MIS 3; 58–24 ka). Taken at face value, the
74 ka age from unit A gives the timing of the preceding
deglaciation, but we cannot rule out the possibility that
the sediment suffers from some incomplete bleaching,
and that the true deposition age is younger.
What is required to make these OSL ages older
(MIS 5)?
To capture uncertainties in the ages conservatively, Fig.
7 and Table 4 show the OSL ages calculated under a
variety of assumptions. Even when considering the
broadest age range for each sample, all upper nine ages
fall within the Middle Weichselian (MIS 3) and no
rea-sonable adjustments in assumptions can collectively
bring their ages to the Early Weichselian (MIS 5 a/c,
474 ka). As the sensitivity analysis showed the
calcu-lated age to depend most on changes in the equivalent
dose and various factors influencing the dose rate (Fig.
8), we tested what those factors would have to be to
obtain
90 ka ages from these samples. To get early
Weichselian ages from the current data we would need
to double the equivalent doses or half the dose rates; the
latter for example by using water contents
4100% by
weight, or a combination of these. We consider it
un-likely that any of these parameters is in error to this
degree.
Comparison with other chronologies and records
A mean age of 44
6 ka (n = 9) for the Pilgrimstad
se-diments agrees with the bulk of previous age
determi-nations from the site, which fall between 60 and 45 ka
0 1 2 3 0 100 200 300 400 500 600 500 1000 1500 2000 2500 3000 20 0 10 30 40 Stimulation time (s) Photon counts / 0.16s Sample 061345
Regenerative dose (Gy)
Sensitivity corrected O
SL
Fig. 6. Single-aliquot regenerative-dose (SAR) OSL growth curve for a single aliquot of sample 061345. The equivalent dose (ED) (open diamond) is obtained by interpolating the corrected natural signal (open triangle) on the growth curve. Repeated dose points (open squares) and a zero dose point (open circle) are also shown. (Inset) The natural decay curve.
Table 4. Pilgrimstad OSL sample results and ages. Listed in order of sample number.
Sample Age (ka) Equivalent dose (Gy) n Dose rate (Gy/ka) Recycling ratio Mean Median Dry1 Saturated2
061344 435 37 364 435 11612 18 2.720.10 1.140.24 061345 454 44 374 464 11810 19 2.620.10 1.060.26 061346 488 49 325 397 13423 18 2.820.09 1.060.31 061347 524 50 424 534 14510 16 2.770.10 1.080.25 061348 494 49 403 504 13910 22 2.830.10 1.030.20 081318 383 36 303 383 1017 28 2.690.10 1.170.09 081319 393 34 293 403 977 31 2.510.09 1.070.12 081320 363 35 272 363 897 32 2.520.09 1.050.17 081321 463 44 373 473 1436 30 3.100.13 1.100.12 081322 745 72 584 755 2029 30 2.740.11 1.070.12 1
Age calculated with water content at time of sampling.
(Fig. 3) (Wohlfarth 2009). These radiocarbon dates are
considered to be of acceptable quality, although close
to the methodological limit (Wohlfarth 2009). The
mean OSL age also agrees with the finite of our two
AMS
14C ages (39.2
2
14C ka BP; Table 5) and is not
contradicted by the non-finite one.
Oxygen isotope curves from the Greenland Ice Sheet
provide estimates of temperature variations during
the Weichselian in the North Atlantic region (e.g.
Johnsen et al. 2001). If the sediments at Pilgrimstad are
of MIS 3 age, as the OSL and other ages suggest,
we would expect a correlation with one of the warmest
and/or longest interstadials during this time – with
duration long enough to allow for the subarctic–arctic
flora and fauna to become established. Taking the
OSL ages at face value leads to a correlation with
Greenland interstadials 12–10 (47–41 ka), but
con-sidering the uncertainties and the actual spread in ages,
interstadials 17–10 are also possible options (cf. Walker
et al. 1999).
This is in agreement with Wohlfarth (2009), who
tentatively places the Pilgrimstad site within Greenland
interstadials 17–12. H ¨attestrand (2008) proposes a
si-milar age shift of the T ¨arend ¨o II Interstadial in
north-ern Sweden, i.e. from the Early Weichselian (MIS 5 a)
to the Middle Weichselian (MIS 3). Our OSL data,
from a critical location in the former central area of the
ice sheet, thus support the documented ice-free
condi-tions reported elsewhere in Fennoscandia during MIS 3
(e.g. Olsen et al. 2001; Arnold et al. 2002; Kjær et al.
2006; Helmens et al. 2007a, b; Ukkonen et al. 2007;
Lunkka et al. 2008; Salonen et al. 2008) (see also Fig. 1).
Change in parameter (%) 30 35 40 45 50 55 –100 –80 –60 –40 –20 0 20 40 60 80 100
Change in age (ka)
Fig. 8. Sensitivity analysis showing by how much the age will change (percentage change) in a given parameter for sample (061)344 (age = 43 ka). Steeper slopes indicate that the age is more sensitive to changes in the given parameter. The age is most sensitive to changes in the equivalent dose, beta source calibration and dose rate, followed by water content; very large changes in these parameters are needed to make the age consistent with MIS 5. The range of x-axis parameter values corresponds to our estimate of 2 SD in that parameter. Similar re-sults were obtained for the other samples. 0 A B 10 20 30 40 50 60 70 80 90 318,H 345,H 320,H 347,H 348,H 344,G 319,G 346,F 321,C 322,A Sample and stratigraphic unit
Agr (ka) 0.6 0.8 1.0 1.2 1.4 1.6 Recycling ratio
Fig. 7. A. Graphical summary of ages for all samples. B. Corresponding recycling ratios. OSL ages are shown in stratigraphic order with youngest ages to the left and oldest to the right. H–A refers to the stratigraphic units in Fig. 3. Age calculated from the population of accepted aliquots using the mean and median, and for natural water content at sampling (dry) and sa-turated water content (range shown by grey square). Error bars represent 1 SD. The SD of mean of ages excludes the older, strati-graphically lower sample (081)322. Note that the water content for organic-rich sample (061)346 has been adjusted to compensate for compression. Marine isotope stages (MIS) in-dicated on the right.
Implications for Weichselian history
Moving the age of the Pilgrimstad Interstadial from
Early to Middle Weichselian (MIS 5 a/c to MIS 3) has
implications for the understanding of glacial history
and environmental change in Scandinavia during the
Weichselian. An ice-free Pilgrimstad at
50–40 ka
re-quires that the Scandinavian Ice Sheet at that time was
restricted, possibly limited to the highest mountains (cf.
Arnold et al. 2002), i.e. much smaller than previously
believed (Lundqvist 2002; Mangerud 2004;
Houmark-Nielsen 2007) (Fig. 1). It also implies that the ice sheet
must have expanded rapidly thereafter to reach
south-eastern Denmark just prior to 30 ka (the Klintholm
ad-vance; Houmark-Nielsen & Kjær 2003; Ukkonen et al.
2007).
Although the vegetation reconstruction for
Pilgrim-stad (Robertsson 1988b) still largely holds true,
recent studies indicate that forest may not have been as
close as suggested (Helmens et al. 2007a). Nevertheless,
a reconciliation of the Pilgrimstad record with
temporaneous records of tundra in northern
con-tinental Europe (Behre 1989) requires alternative
migration routes for vegetation, and/or possible tree
refugia, during previous stadials and different climate
gradients and conditions compared to today (e.g.
Nils-son 1972: p. 225; Ukkonen et al. 2007).
Conclusions
A new exposure at the Pilgrimstad site in central
Sweden shows
44 m thick subtill minerogenic and
organic sediments; these have been OSL-dated.
The OSL ages from the lacustrine and fluvial
sedi-ments range from 52 to 36 ka, while the underlying
glacifluvial deposit is probably younger than
74 ka.
The OSL samples passed appropriate
methodologi-cal checks and the ages are internally consistent. We
therefore consider the results reliable.
From sedimentological and chronological points of
view, we favour deposition during a single but
vari-able event, rather than two separate interstadials.
The OSL ages assign the Pilgrimstad Interstadial
site to MIS 3; the organic deposits could possibly
correspond to one or more of Greenland
inter-stadials 17–10.
The central location of the site, together with results
from other studies in Fennoscandia, indicates that
for at least parts of MIS 3 the ice sheet was absent,
or at least restricted to the highest Scandinavian
mountains.
Acknowledgements. –We thank Jan Lundqvist, Ann-Marie Rober-tsson and Barbara Wohlfarth at Stockholm University for valuable discussions and for sharing experiences and data; Jan-Pieter Buylaert and the technical staff at the Nordic Laboratory for Luminescence Dating for helping with OSL measurements; Damian Steffen at the University of Bern for introducing Alexanderson and Johnsen to the world of de-convolution; and Dick Olsson and Claes in Pilgrimstad for giving us access to the gravel pit and assisting with excavation. The study was financed by a grant (no. 60-1356/2005) from the Geological Survey of Sweden to Alexanderson, with additional fieldwork funding from the Swedish Nuclear Fuel and Waste Management Company (SKB). Alexanderson was in charge of and carried out most of the OSL sampling, preparation and measurements, sedimentological and stratigraphical work as well as writing, with significant input at all stages by Johnsen. Johnsen analysed all the data, with some input from Murray and Alexanderson. Murray contributed to discussions regarding OSL preparation and measurement setup and interpreta-tion of results as well as provided laboratory access. The constructive comments on the manuscript from reviewers Per M ¨oller and Ole Bennike are appreciated.
References
Adrielsson, L. & Alexanderson, H. 2005: Interactions between the Greenland Ice Sheet and the Liverpool Land coastal ice cap during the last two glaciation cycles. Journal of Quaternary Science 20, 269–283.
Alexanderson, H. & Murray, A. S. 2007: Was southern Sweden ice free at 19–25 ka, or were the post LGM glacifluvial sediments in-completely bleached? Quaternary Geochronology 2, 229–236. Alexanderson, H., Johnsen, T., Wohlfarth, B., N ¨aslund, J.-O. &
Stroeven, A. 2008: Applying the optically stimulated luminescence (OSL) technique to date the Weichselian glacial history of southern Sweden. Reports from the Department of Physical Geography and Quaternary Geology, Stockholm University 4, 1–33.
Arnold, N. S., van Andel, T. H. & Valen, V. 2002: Extent and dy-namics of the Scandinavian Ice Sheet during oxygen isotope stage 3 (65,000–25,000 yr B.P.). Quaternary Research 57, 38–48.
Banerjee, D., Murray, A. S., Bøtter-Jensen, L. & Lang, A. 2001: Equivalent dose estimation using a single aliquot of polymineral fine grains. Radiation Measurements 33, 73–94.
Behre, K.-E. 1989: Biostratigraphy of the last glacial period in Eur-ope. Quaternary Science Reviews 8, 25–44.
Behre, K. E. & van der Plicht, J. 1992: Towards an absolute chronol-ogy for the last glacial period in Europe: Radiocarbon dates from Oerel, northern Germany. Vegetation History and Archaeobotany 1, 111–117.
Bøtter-Jensen, L., Bulur, E., Duller, G. A. T. & Murray, A. S. 2000: Advances in luminescence instrument systems. Radiation Measure-ments 32, 523–528.
Choi, J. H., Duller, G. A. T. & Wintle, A. G. 2006: Analysis of quartz LM-OSL curves. Ancient TL 24, 9–20.
Fr ¨odin, G. 1954: De sista skedena av Centralj ¨amtlands glaciala his-toria. Geographica 24, 1–251.
Fuchs, M. & Owen, L. A. 2008: Luminescence dating of glacial and associated sediments: Review, recommendations and future direc-tions. Boreas 37, 636–659.
Garcı´a Ambrosiani, K. 1990: Pleistocene Stratigraphy in Central and Northern Sweden. Ph.D. thesis. Department of Quaternary Re-search, Report 16, 1–15. University of Stockholm.
Hansen, L., Funder, S., Murray, A. S. & Mejdahl, V. 1999: Lumi-nescence dating of the last Weichselian glacier advance in East Greenland. Quaternary Geochronology (Quaternary Science Re-views) 18, 179–190.
H ¨attestrand, M. 2008: Vegetation and Climate During Weichselian Ice Free Intervals in Northern Sweden. Ph.D. thesis. Dissertation from Table 5. 14C ages from twigs.
Sample Stratigraphic unit 14C age years BP
LuS 6957 F 39 2002000
the Department of Physical Geography and Quaternary Geology 15, 1–35. Stockholm University.
Heijnis, H. 1992: Uranium/Thorium Dating of Late Pleistocene Peat Deposits in N.W. Europe. Thesis, Rijksuniversiteit Groningen, 149 pp. Helmens, K. F., Bos, J. A. A., Engels, S., Van Meerbeeck, C. J., Bohncke, S. J. P., Renssen, H., Heiri, O., Brooks, S. J., Sepp ¨a, H., Birks, H. J. B. & Wohlfarth, B. 2007a: Present-day temperatures in northern Scandinavia during the last glaciation. Geology 35, 987–990.
Helmens, K. F., Johansson, P. W., R ¨as ¨anen, M. E., Alexanderson, H. & Eskola, K. O. 2007b: Ice-free intervals at Sokli continuing into Marine Isotope Stage 3 in the central area of the Scandinavian gla-ciations. Geological Society of Finland, Bulletin 79, 17–39. Houmark-Nielsen, M. 2007: Extent and age of Middle and
Late Pleistocene glaciations and periglacial episodes in southern Jylland, Denmark. Bulletin of the Geological Society of Denmark 55, 9–35.
Houmark-Nielsen, M. 2008: Testing OSL failures against a regional Weichselian glaciation chronology from southern Scandinavia. Boreas 37, 660–677.
Houmark-Nielsen, M. & Kjær, K. H. 2003: Southwest Scandinavia, 40–15 kyr BP: Palaeogeography and environmental change. Journal of Quaternary Science 18, 769–786.
Johnsen, S. J., Dahl-Jensen, D., Gundestrup, N., Steffensen, J. P., Clausen, H. B., Miller, H., Masson-Delmotte, V., Sveinbj ¨ornsdo´t-tir, A´. E. & White, J. 2001: Oxygen isotope and palaeotemperature records from six Greenland ice-core stations: Camp Century, Dye-3, GRIP, GISP2, Renland and NorthGRIP. Journal of Quaternary Science 16, 299–307.
Kjær, K. H., Lagerlund, E., Adrielsson, L., Thomas, P. J., Murray, A. & Sandgren, P. 2006: The first independent chronology for Middle and Late Weichselian sediments from southern Sweden and the island of Bornholm. Geologiska F ¨oreningens; Stockholm F ¨orhandlingar 128, 209–220.
Kulling, O. 1945: Om fynd av mammut vid Pilgrimstad i J ¨amtland. Geological Survey of Sweden C 473, 61 pp.
Kulling, O. 1967: Yttrande med anledning av J. Lundqvists f ¨oredrag ‘Submor ¨ana sediment i J ¨amtland’. Geologiska F ¨oreningens i Stock-holm F ¨orhandlingar 89, 123–125.
Lagerb ¨ack, R. 2007: Ventifacts – means to reconstruct glacial develop-ment and palaeoenvirondevelop-ment in northern and central Sweden. Geo-logiska F ¨oreningens; Stockholm F ¨orhandlingar 129, 315–324. Larsen, E. & Mangerud, J. 1992: Subglacially Formed Clastic Dikes.
Geological Survey of Sweden, SGU Ca 81, 163–170.
Linde´n, M., M ¨oller, P. & Adrielsson, L. 2008: Ribbed moraine formed by subglacial folding, thrust stacking and lee-side cavity infill. Boreas 37, 102–131.
Lundqvist, J. 1967: Submor ¨ana sediment i J ¨amtlands l ¨an. Geological Survey of Sweden C 618. 267 pp.
Lundqvist, J. 2002: Weichseltidens huvudfas. In Frede´n, C. (ed.): Berg och jord, 124–135. Sveriges Nationalatlas, V ¨allingby.
Lundqvist, J. & Miller, U. 1992: Weichselian Stratigraphy and Glacia-tions in the T ˚asj ¨o–Hoting Area, Central Sweden. Geological Survey of Sweden C 826, 35 pp.
Lunkka, J. P., Murray, A. & Korpela, K. 2008: Weichselian sediment succession at Ruunaa, Finland, indicating a Mid-Weichselian ice-free interval in eastern Fennoscandia. Boreas 37, 234–244. Mangerud, J. 2004: Ice-sheet limits on Norway and the Norwegian
continental shelf. In Ehlers, J. & Gibbard, P. L. (eds.): Quaternary Glaciations – Extent and Chronology, 271–294. Elsevier, Am-sterdam.
Murray, A. S. & Wintle, A. G. 2000: Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Ra-diation Measurements 32, 57–73.
Murray, A. S. & Wintle, A. G. 2003: The single aliquot regenerative dose protocol: Potential for improvements in reliability. Radiation Measurements 37, 377–381.
Murray, A. S., Marten, R., Johnson, A. & Martin, P. 1987: Analysis for naturally occurring radionuclides at environmental concentra-tions by gamma spectrometry. Journal of Radioanalytical and Nu-clear Chemistry Articles 115, 263–288.
Nilsson, T. 1972: The Pleistocen. 508 pp. Esselte Studium, Stockholm. Olsen, L., Sveian, H. & Bergstrøm, B. 2001: Rapid adjustments of the western part of the Scandinavian Ice Sheet during the Mid and Late Weichselian – a new model. Norsk Geologisk Tidsskrift – Norwegian Journal of Geology 81, 93–118.
Prescott, J. R. & Hutton, J. T. 1994: Cosmic ray contributions to dose rates for luminescence and ESR dating: Large depths and long-term time variations. Radiation Measurements 23, 497–500. Preusser, F., Andersen, B. G., Denton, G. H. & Schl ¨uchter, C. 2005:
Luminescence chronology of Late Pleistocene glacial deposits in North Westland, New Zealand. Quaternary Science Reviews 24, 2207–2227.
Robertsson, A.-M. 1986: The diatom flora of the till-covered sedi-ments at Pilgrimstad, J ¨amtland. Striae 24, 2123–2220.
Robertsson, A.-M. 1988a: Biostratigraphical Studies of Interglacial and Interstadial Deposits in Sweden. Ph.D. thesis. Department of Quaternary Research, Report 10, 1–19. University of Stockholm. Robertsson, A.-M. 1988b: A reinvestigation of the biostratigraphy of
the till-covered sediments at Pilgrimstad and O¨je, central Sweden. In Robertsson, A.-M. (ed.): Biostratigraphical Studies of Inter-glacial and Interstadial Deposits in Sweden, 1–35. University of Stockholm, Department of Quaternary Research, Stockholm. Robertsson, A.-M. & Garcı´a Ambrosiani, K. 1992: The Pleistocene in
Sweden – a Review of Research, 1960–1990. Geological Survey of Sweden Ca 81, 299–306.
Salonen, V.-P., Kaakinen, A., Kultti, S., Miettinen, A., Eskola, K. O. & Lunkka, J. P. 2008: Middle Weichselian glacial event in the cen-tral part of the Scandinavian Ice Sheet recorded in the Hitura pit, Ostrobothnia, Finland. Boreas 37, 38–54.
Spencer, J. Q. & Owen, L. A. 2004: Optically stimulated luminescence dating of Late Quaternary glaciogenic sediments in the upper Hunza valley: Validating the timing of glaciation and assessing dating methods. Quaternary Science Reviews 23, 175–191. Svendsen, J. I., Alexanderson, H., Astakhov, V. I., Demidov, I.,
Dowdeswell, J. A., Funder, S., Gataullin, V., Henriksen, M., Hjort, C., Houmark-Nielsen, M., Hubberten, H. W., Ingo´lfsson, O´., Jakobsson, M., Kjær, K. H., Larsen, E., Lokrantz, H., Lunkka, J. P., Lys ˚a, A., Mangerud, J., Matiouchkov, A., Murray, A., M ¨oller, P., Niessen, F., Nikolskaya, O., Polyak, L., Saarnisto, M., Siegert, C., Siegert, M. J., Spielhagen, R. F. & Stein, R. 2004: Late Qua-ternary ice sheet history of northern Eurasia. QuaQua-ternary Science Reviews 23, 1229–1271.
Thomas, P. J., Murray, A. S., Kjær, K. H., Funder, S. & Larsen, E. 2006: Optically stimulated luminescence (OSL) dating of glacial sediments from Arctic Russia – depositional bleaching and metho-dological aspects. Boreas 35, 587–599.
Ukkonen, P., Arppe, L., Houmark-Nielsen, M., Kjær, K. H. & Karhu, J. A. 2007: MIS 3 mammoth remains from Sweden – implications for faunal history, palaeoclimate and glaciation chronology. Quaternary Science Reviews 26, 3081–3098.
Walker, M. J. C., Bj ¨orck, S., Lowe, J. J., Cwynar, L. C., Johnsen, S., Knudsen, K.-L. & Wohlfarth, B. & INTIMATE group 1999: Iso-topic ‘events’ in the GRIP ice core: A stratotype for the Late Pleis-tocene. Quaternary Science Reviews 18, 1143–1150.
Wohlfarth, B. 2009: Ice-free conditions in Fennoscandia during Marine Oxygen Isotope Stage 3? Swedish Nuclear Fuel and Waste Management Company (SKB) Report TR-09-12, 1–48.
Late Quaternary ice sheet history
and dynamics in central and
southern Scandinavia
Timothy F. Johnsen
Doctoral Thesis in Quaternary Geology at Stockholm University, Sweden 2010
linked to our understanding of how ice sheets have operated in the
past. Recent work suggests an emerging new paradigm for the
Scan-dinavian ice sheet (SIS); one of a dynamically fluctuating ice sheet.
This doctoral research project explicitly examines the history and
dynamics of the SIS at four sites within Sweden and Norway, and
provides results covering different time periods of glacial history.
Two relatively new dating techniques are used to constrain the ice
sheet history.
Dating of sub-till sediments in central Sweden and central Norway
indicate ice-free conditions during times when it was previously
inferred the sites were occupied by the SIS. Consistent exposure
ages of boulders from the Vimmerby moraine in southern Sweden
indicate that the southern margin of the SIS was at the Vimmerby
moraine ~14 kyr ago. In central Sweden, consistent exposure ages
for boulders at high elevation agree with previous estimates for the
timing of deglaciation around 10 ka ago, and indicate rapid thinning
of the SIS during deglaciation.
Altogether this research conducted in different areas, covering
dif-ferent time periods, and using comparative geochronological
met-hods demonstrates that the SIS was highly dynamic and sensitive to
environmental change.
Department of Physical Geography
and Quaternary Geology
ISBN 978-91-7447-068-0 ISSN 1653-7211
I was born and raised on Vancouver Island on the west coast of Canada surrounded by beautiful mountains and coastline, where I developed a deep curiosity and passion for understanding the workings of nature. I completed a Bachelor of Science degree with distinction in Geography 1998 at the University of Victoria, Canada. Then I completed a Masters of Science degree in Geography 2004 at Simon Fraser University, Canada, for which I was awarded the Canadian Association of Geographers Starkey-Robinson Award 2005. I began a PhD in 2004 in Stockholm, Sweden investigating the dynamics of the Scandinavian ice sheet, and eating brown cheese with waffles under the midnight sun.